Diabetes Mellitus and Cardiovascular Disease
Patients with type 2 diabetes mellitus (DM) have an increased risk of atherosclerotic disease, including coronary heart disease, cardiovascular disease, and peripheral vascular disease. Diabetes itself1 and not just the associated risk factors of dyslipidemia, hypertension, and obesity contributes a major portion of this risk. In particular, the level of hyperglycemia may play a key role. While the relationship of increased blood glucose to microvascular complications is well-recognized7-9, its relation to atherogenesis was, until recently, less well documented. A prospective, population-based study in middle-aged and elderly patients in Finland with type 2 DM has shown a linear correlation between baseline fasting blood glucose (FBG), or HbA1c, and coronary heart disease mortality10. In the WESDR database, subjects diagnosed with diabetes at age 30 years or older had a statistically significant increase in mortality from vascular causes for every 1% increase in glycosylated hemoglobin, with a hazard ratio of 1.10 to 1.28 for various types of events11. The Islington Diabetes Survey found a linear association between 2-hour postprandial glucose or HbA1c and coronary heart disease, with the stronger association with the 2-hour glucose test12. In the San Antonio Heart Study, the level of hyperglycemia was a strong, independent predictor of all-cause and cardiovascular mortality13.
A growing body of evidence indicates that the increased risk for macrovascular complications associated with type 2 DM extends to patients with glucose abnormalities that do not meet the criteria for frank diabetes. The Hoorn study found an increased risk of all-cause and cardiovascular mortality with higher 2-hour post-load glucose values and increasing HbA1c in a non-diabetic general population of men and women15. In the EPIC study, an increase of 1% in HbA1c was associated with a 28% increase in risk of death and an increase of approximately 40% in cardiovascular or coronary heart disease mortality in a cohort of 4,662 men4. Although diabetic subjects were included in this trial, and diabetes was found to be an independent predictor of cardiovascular risk when evaluated separately from HbA1c (another independent predictor), only HbA1c and not diabetes predicted CV death when both were included in the same analysis, further strengthening the link between glucose elevations and CV risk, versus the presence or absence of diabetes. Similarly, a study in non-diabetic elderly women found that all-cause mortality and coronary heart disease were significantly related to fasting plasma glucose16.
In a study from Oslo17, non-diabetic men aged 40-59 years had a significantly higher cardiovascular mortality rate if their FPG was >85 mg/dL. Long-term follow-up of several prospective European cohort studies has confirmed a higher risk of cardiovascular-related mortality in non-diabetic men with the highest 2.5% of values of FPG and 2-hour postprandial glucose18. A meta-regression analysis of data from 20 cohort studies found a progressive relationship between glucose levels and cardiovascular risk even below the cutoff points for diagnosis of DM3. Likewise, in the 23-year Paris Prospective Study19 of 7,018 middle-aged nondiabetic men, increased fasting or 2-hour postprandial blood glucose was associated with increased total and coronary mortality in a graded, non-threshold relationship.
The American Diabetes Association (ADA) has recognized an intermediate category of IFG, defined as a fasting plasma glucose of 6.1-6.9 mmol/L (110-125 mg/dL)6, as well as the older category of IGT, defined as a 2-hour glucose level of 7.8-11.1 mmol/L (140-199 mg/dL) after a 75 gram oral glucose load, with FPG levels below 7.0 mmol/L. The European Diabetes Epidemiology Group, based on a meta-analysis of 10 prospective European cohort studies, found that IGT was associated with survival curves intermediate between those of non-diabetic and diabetic subjects, while IFG curves were similar to those of normoglycemic subjects. A direct comparison revealed that IFG had a higher specificity (79%) for predicting cardiovascular disease than IGT (57%), but IGT was a more sensitive (54%) predictor than IFG (28%) in predicting incident cardiovascular disease20.
In summary, the data cited above demonstrate that people with IFG and IGT (collectively referred to as “prediabetes”) have an excess risk of development of overt type 2 diabetes, coronary heart disease, cerebrovascular disease, and peripheral vascular disease compared to a population with normal fasting and 2-hour postprandial glucose levels. Further, a continuum of increasing risk appears to exist, as opposed to a threshold level of hyperglycemia below which no increased risk prevails2-4. IGT and IFG subjects are currently unlikely to receive glucose-lowering treatment with existing pharmacotherapies. Their under-treated dysglycemia represents a large unmet medical need, and a large public health issue.
A number of large intervention studies have been conducted over the last two decades in both type 1 and type 2 diabetic patients. The primary aim of these trials was to evaluate the impact of improved metabolic control on microvascular endpoints and the studies were designed and sized accordingly.
Macrovascular outcomes were included in these trials as secondary endpoints and although the treatment differences seen were not statistically significant, trends were evident in each trial of an association between intensified glycemic control and reduced cardiovascular mortality and morbidity.
The two principal intervention trials in recent years were the Diabetes Control and Complications Trial (DCCT) in type 1 diabetic patients7 and the United Kingdom Prospective Diabetes Study (UKPDS) in type 2 DM14. In the DCCT, cardiovascular events decreased by 41% in the intensively-treated group, but this difference was not statistically significant. In the UKPDS, which compared the effects of intensive management to the effects of standard care on micro- and macrovascular complications in 3,642 type 2 diabetic subjects followed for a median of 10.4 years, a significant decrease in microvascular complications was observed in the intensive treatment group, which achieved a significantly lower median HbA1c of 7.0% compared to the standard group (median HbA1c 7.9%). Although strongly suggestive, the intervention data from this study failed to show a statistically significant decrease in the endpoint of myocardial infarction, which decreased by 16% with the 0.9% decrease in HbA1c (p=0.052). However, epidemiologic analysis of the UKPDS database2 revealed that a single percent point decrease in HbA1c was associated with a 25% reduction in diabetes-related death, a 7% reduction in all-cause mortality, and an 18% reduction in fatal and nonfatal MI. Similar reductions in the risk for stroke, amputation and congestive heart failure were seen with decreasing HbA1c. These associations of HbA1c with cardiovascular risk were without threshold, i.e. they occurred across the entire study population.
In the 8-year Kumamoto study3 of intensive multiple-dose insulin treatment of type 2 diabetic patients, half as many serious macrovascular events (MI, angina, stroke, claudication, gangrene, or amputation) occurred in the intensive treatment arm as in the conventional treatment arm. This reduction was not statistically significant, in all likelihood because of the small size of the trial (n=110). Several large prospective trials, including the ACCORD trial (NHLBI) and the VA diabetes trial, are now ongoing or planned to specifically and primarily evaluate the hypothesis that treatment of diabetes in patients with cardiovascular risk factors will reduce cardiovascular morbidity and mortality.
Recent intervention studies in IGT have focused on the reduction of rates of progression to type 2 diabetes. Lifestyle interventions (primarily institution of diet and exercise plans) have led to striking reductions in progression from IGT to diabetes in both the recently-completed NIH-sponsored DPP in North America and the Finnish DPS lifestyle study. Each trial was terminated early after independently demonstrating a 58% reduction versus controls in development of new cases of type 2 DM from IGT in the lifestyle intervention arm21,33. Lifestyle changes were pursued aggressively in both studies, and whether such interventions can be maintained indefinitely is an open question.
Pharmacotherapies have also been tested in delaying the development of type 2 DM:                Metformin treatment of IGT in the DPP study was associated with a statistically significant 31% reduction in the rate of progression to type 2 DM.21         Acarbose in the STOP-NIDDM trial reduced the progression from IGT to type 2 DM from 41.8% in the placebo arm to 32.8% over 3.6 years' median duration of treatment (p<0.05) as well as reducing the risk of CV events by 49%.40         Troglitazone in the halted TRIPOD study; 12.3% of placebo-treated subjects versus 5.4% of troglitazone-treated subjects with prior gestational diabetes developed type 2 DM over a mean of 30 months of treatment (p<0.05).45         
With the exception of the STOP-NIDDM study, cardiovascular risk reduction data from these recent diabetes prevention studies are all pending publication. At present the only other data available on CV risk reduction in the IGT/IFG population from treatment with pharmacologic antihyperglycemic agents come from a small Swedish study conducted in the 1960s which demonstrated a reduction in CV events in IGT subjects with the use of tolbutamide22,23. Clearly new therapies for glucose lowering must be tested for their effects on serious cardiovascular outcomes in this population.
Recent evidence has provided support for a beneficial effect of insulin treatment in improving cardiovascular outcomes in patients with diabetes. The DIGAMI study24, in which diabetic patients hospitalized with acute MI were allocated to receive an IV insulin-glucose infusion in-hospital followed by intensive chronic outpatient treatment with insulin, versus standard treatment, showed a significant 28% reduction of all-cause mortality in the patients who received intensive insulin treatment. Most of these deaths were cardiovascular in etiology. The most striking reductions in mortality were seen in the subset of patients without prior insulin treatment, with low cardiovascular risk pre-MI. In those subjects significant survival differences were even seen pre-discharge, while still in hospital post-MI, and enhanced survival in the same cohort was also seen in long-term followup.
Part of the benefits of insulin treatment was likely due to improved long-term glycemia post-MI, but the in-hospital results suggest that other, more acute, effects of insulin besides long-term glycemic control may have played a role, such as improved platelet function, decreased PAI-I levels, and insulin-mediated reductions in circulating free fatty acid levels with consequent improved dyslipidemia and decreased myocardial oxygen requirement. Chronic insulin therapy may thus provide a level of protection against the cumulative deleterious effect of even subacute episodes of ischemia, and on the progression of atherosclerosis.
A recent study from Belgium25 reinforces the beneficial role of insulin treatment of critically-ill subjects. In this trial, critical-care post-surgical patients with random blood glucose values greater than 110 mg/dL were treated while in ICU either with an insulin infusion to lower blood glucose to the 80-110 mg/dL range (intervention); or to receive insulin infusions only if blood glucose exceeded 215 mg/dL, with the aim of infusion to reduce blood glucose to between 180 and 200 mg/dL (control). Twelve-month follow-up showed significantly different reductions of 8.0% and 4.6% in overall mortality in the intervention and control groups respectively, and most of the benefit was attributable to the cohort of subjects who were in ICU for 5 days or more. In-hospital mortality, septicemia, acute renal failure and hemodialysis incidence, and transfusion requirements were also significantly reduced in the intervention group versus the control group.
The use of exogenous insulin in a IGT, IFG, or diabetic population should confer several potential metabolic and cardiovascular benefits associated with insulin treatment:    1. A powerful effect to delay the exposure of target tissues to toxic levels of glycemia that is finely titratable and durable, compared to oral antidiabetic agents.    2. Suppression of circulating free fatty acids (FFA) with:            Reduced VLDL synthesis and improved lipoprotein patterns (lower triglycerides, increased HDL-C)        Reduced lipotoxicity at the level of the beta cell and on insulin's target tissues        Reduced obligatory oxidative metabolism in ischemic myocardium            3. Prevention of metabolic decompensation (including both glucose and FFA) due to stress, both mild and frequent (daily stresses and minor illness or injury) and severe and less common (major injury, illness, surgery, vascular events). These stress events will suppress endogenous insulin responses even when a pharmacologic secretagogue or sensitizer is present, but exogenous, injected insulin cannot be so suppressed.    4. In addition, recent work has demonstrated direct associations between insulin treatment and enhanced nitric oxide-mediated vasodilatation, which is impaired in insulin-resistant states such as IGT, IFG and diabetes26,27. Moreover, reductions in the endothelial dysfunction28 and inflammation29 that are characteristic of both diabetes and atherosclerosis have been demonstrated following insulin treatment.
Whereas insulin therapy is undoubtedly efficacious in reducing blood glucose concentrations and, as noted above, may hypothetically improve survival in individuals with dysglycemia, outcome studies using insulin in this population have not been done to date. Several reasons may account for this including a) the need for insulin to be injected as opposed to be taken orally; b) concerns regarding the side effect of hypoglycemia; (low blood glucose) c) epidemiologic evidence linking high serum insulin levels to macrovascular disease; d) the very recent recognition that glucose is a risk factor for cardiovascular outcomes across the range from normal through all stages of diabetes; e) lack of predictability in the action of long-acting insulins; and f) lack of experience in achieving near-normal glucose levels with insulin preparations available to date.
Many of these issues are, however, not relevant today. First, it is now widely recognized that the epidemiologic relationship between hyperinsulinemia and macrovascular disease is extremely unlikely to imply a cause-effect relationship. This is based on randomized controlled trial evidence from both the DCCT and the UKPDS trials that individuals who were given exogenous insulin in an effort to reduce the risk of microvascular disease had a trend towards fewer, not more, adverse cardiovascular outcomes. This conclusion is supported by other studies discussed above, including the DIGAMI study, the Kumamoto study, a meta-analysis of studies of intensified insulin therapy in type 1 diabetes, and several other analyses. It therefore appears that hyperinsulinemia as a result of exogenous administration of insulin is not a cardiovascular risk factor. Second, the potential of intensified insulin therapy has achieved new attention in light of the UKPDS and the potential benefits of tight glycemic control in people with type 2 diabetes. Third, the simplicity of glucose monitoring devices and the decreasing costs of home glucose monitoring, as well as the negligible discomfort associated with today's injection devices available today, have made injections and blood glucose monitoring more accessible and easier for patients to accomplish. Fourth, there is growing recognition of the importance of metabolic abnormalities as a cause of cardiovascular disease. Finally, there are new long acting analog insulins having properties such as a longer duration of action (up to 24 hours) and a smoother profile, with a less defined peak of action which make them viable treatment agents in the IGT, IFG early Type 2 diabetes populations.
Lantus® LANTUS (insulin glargine) is a recombinant human insulin analog that is a long-acting (up to 24-hour duration of action), parenteral blood-glucose lowering agent.39 The post-marketing surveillance safety database experience reveals no increased incidence of hypoglycemia or unexpected adverse reactions compared to other marketed insulin preparations. In a multiple-dose pharmacokinetic study, Lantus® LANTUS (insulin glargine) levels were shown to reach steady-state after 2 doses (2 days) of treatment (Study 1020). Treatment with Lantus® insulin LANTUS (insulin glargine) offers the possibility of a smooth, daylong, blood insulin profile without a definite peak that can be finely titrated to lower subjects' FPG in a durable manner, while minimizing the risk of hypoglycemia at other times of day.
However, a central question concerns the administration of insulin to nondiabetic or early diabetic subjects and the propensity for hypoglycemia this may confer. Insulin has traditionally been reserved for treatment of more severe hyperglycemia, in established type 1, or advanced type 2, diabetes. In these patients the risk of hypoglycemia is greater the closer the achieved blood glucose is to normal.42 
For type 2 diabetic patients, as well as individuals with prediabetes, medical management begins with diet restriction and exercise as tolerated30,31. Even if pharmacotherapy in the form of oral antidiabetic drugs or insulin is needed later, diet and exercise are always the cornerstone of disease management. No drugs are currently approved for the treatment of prediabetes, but most of these individuals are overweight or obese, and successful lifestyle intervention has been shown to improve blood glucose levels and even delay progression to diabetes32,33. Exercise increases blood glucose uptake in muscle, and leads to a reduction in endogenous insulin output, as little insulin is needed to provide fuel to these tissues during exercise.34 Exogenous, pharmacologically-provided insulin present in the circulation cannot be so modulated, and its presence can predispose to hypoglycemia.
Exercise-induced hypoglycemia in insulin-treated diabetes patients is well-described35,41, and is often dealt with in practice by reducing the mealtime insulin dose, or giving oral calories, prior to an exercise session.36,37 These methods are cumbersome, and hypoglycemia is still a risk following exercise. The insulin dose that is most frequently modulated in response to upcoming exercise is the short-acting insulin given before the preceding meal, because these insulins have prominent peaks in their actions, used to target the blood glucose rise that occurs following a meal, but they place patients at increased risk for hypoglycemia if there is a mismatch between insulin availability and calories absorbed. Thus hypoglycemia is a risk e for all insulin-treated individuals, and this risk is enhanced when these individuals exercise, and the closer to normoglycemia they are treated.
The ideal basal insulin might be expected to be less worrisome from this standpoint because the circulating insulin produced would target blood glucose elevations throughout the day rather than mealtime fluctuations. It would not demonstrate notable peaks in plasma level, and in consequence the tendency to produce hypoglycemia would be less than with peaked insulins. The “Treat-to-Target” study38 in US/Canada type 2 diabetic patients investigated whether a single bedtime dose of Lantus® LANTUS (insulin glargine) vs NPH insulin (a moderate- to long-acting insulin with a pronounced peak in plasma activity for 4 to 8 hours after injection)39 would achieve target metabolic control without increasing nocturnal hypoglycemia. The trial was successful in demonstrating both its primary objective (more Lantus® LANTUS (insulin glargine)-treated patients than NPH-treated patients reaching target HbAlc [<=7%] without nocturnal hypoglycemia), but also showed significant reductions in nocturnal hypoglycemia vs NPH in all patients.
If peaked insulins pose a danger for hypoglycemia in advanced type 2 diabetes, they certainly do in milder forms of diabetes, and in prediabetes, where the blood glucose concentrations are only modestly elevated, especially in relation to exercise. Treatment with an insulin with notable peak effects runs a great risk of producing low blood glucose levels that will be bothersome and dangerous to people with these conditions. There exists an unmet medical need to provide insulin treatment to to individuals with milder glucose intolerance who are at high risk for CV disease. Cardiovascular disease in subjects with IGT, IFG, and early diabetes is prevalent and life-threatening. Advances have been made in recent years in treating the associated cardiovascular risk factors of hypertension and hyperlipidemia. Depended upon the results of the morbidity/mortality study association between blood glucose elevations and cardiovascular risk in these subjects is likewise continuous and progressive, treatment of this dysglycemia becomes urgent.
Insulin treatment has been demonstrated to reduce CV morbidity and mortality in a population with more advanced diabetes, and offers this prediabetic population the possibility of reducing cardiovascular risk through effective reductions in blood glucose and free fatty acid levels, and in the associated tissue damage resulting from their chronic elevations. The availability of Lantus® insulin LANTUS (insulin glargine) creates the possibility of treating subjects with widely-varying degrees of dysglycemia with the effectiveness of insulin over a 24 hour period while minimizing the risk of hypoglycemia (especially hypoglycemia seen in association with exercise) inherent in earlier insulin preparations with more distinct peak effects.
Diabetic dyslipidemia (DDL) in type 2 diabetes is another condition where there exists an unmet medical need. DDL is characterized by fasting hypertriglyceridemia, low HDL cholesterol (HDL-C), small dense (atherogenic) LDL particles, and elevated free fatty acid (FFA) concentrations. Whereas lipid disorders associated with type 1 diabetes (hypertriglyceridemia with low LDL) are simpler in etiology, and relate to insulin deficiency which, when replaced, normalizes the plasma lipid profile, the pathophysiology of lipid disturbances in type 2 diabetes is more complex, being partly related to concomitant obesity and insulin resistance. Key factors in the development of lipid abnormalities in type 2 diabetes include:                Failure of suppression of hormone-sensitive lipase in adipose tissue, which leads to increased lipolysis and increased supply of FFA from the adipocyte for, among other things, VLDL-triglyceride (TG) synthesis by the liver        Reduced catabolism of TG-rich particles (such as VLDL), and reduced transfer of surface components of those particles to HDL, partly accounting for the low HDL-C levels seen in DDL.        Accelerated transfer of cholesterol from HDL to other lipoproteins, also contributing to low HDL-C        Reduced clearance of chylomicrons and more atherogenic chylomicron remnants, as well as reduced clearance of other remnant particles (intermediate-density lipoproteins or IDL)        Decreased activity of lipoprotein lipase (LPL) and hepatic TG lipase (HTGL) which break TG down into FFA for fuel in muscle and fat cells.        Overproduction of VLDL by the liver, exacerbated by elevations in glucose and FFA        
Although the lipid abnormalities of type 2 diabetes are more resistant to normalization with antidiabetic treatment, even when that treatment is successful, increased insulinization has been shown to improve most of the defects above, namely, improved lipase activity with reduced lipolysis; increased clearance of chylomicrons; reduced production of VLDL from the liver, both through reduction of FFA substrate and by independent mechanisms; and increases in HDL, generally seen in association with increased LPL activity.
The treatment initially recommended for type 2 diabetes, and reinforced as the cornerstone of management even after pharmacologic treatment is initiated, is diet control and regular exercise. When these lifestyle measures are no longer successful alone in controlling blood glucose levels, pharmacologic treatment is begun, traditionally using oral antidiabetic drugs alone and in combination. Whereas there is no a priori reason why insulin cannot be used to manage mildly diabetic patients, it is usually reserved for late-stage diabetes management because:                Insulin must be given by injection and many patients find injections objectionable        Insulin and injections have acquired the stigma of late-stage management—“if I'm taking insulin my diabetes must be very severe”—and to forestall insulin is a way of saying “my diabetes isn't so bad yet”        
In fact insulin injections have become almost painless in recent years due to improved delivery systems. The “late-stage” stigma is based on tradition and former practice more than any real reason why insulin should be reserved for the end game of diabetes.
The one valid reason for not using insulin in patients as first pharmacotherapy is a concern over the one principal side effect of insulin—low blood glucose, or hypoglycemia. This is an important concern in using insulin to treat early type 2 diabetes primarily because most available insulins have a peak in their plasma activity at a certain time following injection. It is at these times of peak activity that the insulin-treated patient with diabetes is most vulnerable to hypoglycemia, and diets and exercise patterns must often be tailored around the prescribed insulin regimen to avoid hypoglycemia. This risk is greater the closer patients' blood glucose levels come to normal—and yet normoglycemia is the goal of diabetes management.
There is evidence that the scientific community is taking the abnormalities of DDL more seriously than it has in the past in terms of the risk they pose for atherogenesis. The Adult Treatment Panel of the NCEP on the “detection, evaluation, and treatment of cholesterol disorders in adults” authored an update of the ATP II summary in the Fall of 2002. The ATP III took hypertriglyceridemia more seriously than the predecessor ATP II Committee had as a marker for increased CV risk. The ATP III acknowledged that more recent studies, and additional analyses of older studies, have shown elevated TGs to be an independent risk factor for CHD, whereas in the past the association between TG and CHD was not independent from other confounding risk factors such as LDL and HDL abnormalities. ATP III reduced the TG concentration threshold for each degree of abnormality (normal, borderline, high, and very high) from their ATP-II levels, and offered VLDL cholesterol, and serum TG, as markers for atherogenic remnant lipoproteins, which the committee identified as a target for intervention as well as LDL-C. The committee formalized the concept of “non-HDL cholesterol” (non-HDL-C) as a target for therapy in persons with hypertriglyceridemia, perhaps more relevant than LDL-C alone in these individuals. Non-HDL-C was seen as an acceptable surrogate for apo-B in routine clinical practice.
ATP III pointed out that when fasting TG are less than 200 mg/dL, VLDL-C is not markedly elevated, and non-HDL-C correlates very well with LDL-C. As TG rises above 200 mg/dL, the relation between LDL-C and non-HDL-C is looser, and LDL-C alone “inadequately describes the CV risk associated with atherogenic lipoproteins.” When fasting TG exceed 500 m/dL, much of the cholesterol resides in nonatherogenic forms of TG-rich lipoproteins, and non-HDL-C becomes “less reliable as a predictor of CHD risk.” On the other hand, the risk of markedly elevated TG (>500 mg/dL) for pancreatitis has long been recognized, even by FDA, and here too the ability of insulin to reduce these elevations may exceed what other OADs can deliver. Thus there are two categories of elevations in fasting TG that may be amenable to insulin treatment, and for which insulin may be superior to OADs. One is elevations in the 500-1000 mg/dL range, for which hypertriglyceridemia alone is the target, being a surrogate for reduction in risk for pancreatitis. The other is elevations in the 200-500 mg/dL range, for which hypertriglyceridemia is one of a host of biomarkers associated with CHD risk; non-HDL-C, HDL-C, and remnant lipoproteins being other, perhaps more important endpoints in this regard.